CharonY

Yes and the point in these case is of course that while neither of these models predict the next set of numbers, they do tell us something about the state of the system (e.g. if it is loaded or not). Likewise, deviations from expected model behaviour in climate patterns tells us about missing elements and/or relevance of the components incorporated into the model.

Indeed. In fact models could be used to test whether a lottery system deviates from the assumed equivalency between draws, and depending on the type of study this is what some have been doing. E.g. assuming that the the anthropogenic influence of CO2 was not there, what would be the expected temperature trend?

Also, this is a blatant case where there is simply a gross misunderstanding what these models are and what they do. In the case of lottery the model is quite clear. A given ball has the same likelihood getting picked (well, 1/n-i for each pick) thus the model will allow us to test deviations from it. It does not predict the the next ball. That, in theory, could only be achieved if one could accurately model the mechanical movements in a system, in which each state is dependent on the previous one. Thus, beyond a few movements a mechanical simulation trying to predict which ball is going to end up is going to be difficult to near impossible. The same thing is true for weather, for example. Depending on how chaotic your system is in a given geographic area, the ability to somehow predict the weather beyond a few days at most is very, very low. However, climate takes aggregate information over long time frames. I.e. in your example you are basically conflating weather forecast with climate change models. You are essentially stating that because we are unable to predict the precise temperature on July the 23rd in 2041 at 11 AM we cannot assess general patterns. Yet even without models we have little issue in predicting that winters are generally going to be colder than summers. Why, because we look at much rougher patterns. Individual days may vary, but the trend is clear.

That is only true if one approaches evolutionary theories on a very high level in which details are pretty much not considered. Below that level there have been many aspects that have fallen out of favour over time. I would not know what to make out of that comparison. There are many aspects in Chemistry and Physics that are not well understood. Otherwise there would not be ongoing research in those disciplines. The main difference between the disciplines is roughly speaking the complexity of the systems under investigation which lends itself to the need for different methodologies.

While not wrong this specifically refers one specific mechanism, namely natural selection. While this was probably the first recognized mechanism, it is not the only one. In fact I think looking at the expected outcome makes it easier to argue why evolution is the expected outcome. Specifically when we talk about evolution we mean that the genetic composition of a given population is not static. The reason are those that you mentioned, but can also include stochastic effects. Especially in small population random elimination of individuals from the pool, regardless of their genetic composition (and thus, inheritable aspect of reproductive success) can shift the gene pool significantly. As such there are only few situations where you do not expect evolution to happen (e.g. the population being in a Hardy-Weinberg equlibrium).

The point is not really to show feasibility but simply to repeat a talking point to make it sound plausible. Even if you talk about it just to show how improbable or implausible it is, the fact that folks are talking about it makes it more likely that folks (mis-)remember it as a thing. This is a similar principle on which advertisements work. Only if you critically engage with the subject does a discussion or debunking actually has any effect.

As the name implies, FFI is hereditary disease. Insomnia is just one of the symptoms and not the cause of morbidity itself. You will have to stop the prion misfolding in order to avoid fatal nerve damage. I.e. it is not the lack of sleep that leads to death, it is the prion-induced damage (see also Creutzfeldt-Jacob or Kuru for similar diseases).

As someone who has been in working very toxic environments there are a couple of comments I have. But first, let me preface it by stating that obviously by only having a part and only one side of the story I will have to make many assumptions. The hard aspects that are telltale signs of toxic situations are when the boss humiliates their students or postdoc. The first thing of note is of course not to get into that situation in the first place. Abusive PIs have to be avoided and I would understand if as a novice you will have a hard time navigating that. One thing in your description that strikes me is that you describe how folks have behaved toward you, but you present yourself entirely in the passive. I am obviously not asking you to describe things in detail. But it would be important to reflect on how yo communicated with whom and how it may have appeared to them. For example, how did yo manage the apparent conflict between the postdoc and PI? To provide perspective, I was often in situations where I was swooping in to rescue a failing project. What I tried to do was building win-win situations and discourage the PI from assigning blame. I.e. ensure and communicate early on that one wants to go for shared authorships etc. Obviously this is a bit easier as postdoc or perhaps even as senior grad student but trying to build a collaborative atmosphere, even if the PI poisons the atmosphere can mitigate a number of issues. The other half is of course managing the PI's expectations. Especially with regard to changing demands, it is useful to communicate a part in writing. E.g. if in a meeting some changes to a protocol are made, you go back, write up the changes (i.e. not only the data) and send it off to have the PI validate them. That way you have an audit trail of sorts that would show the changes but also would make sure that you may not have misunderstood anything. Or to put it differently, even as a student you can and you should be proactive in managing interactions and expectations but also reflect on how your behaviour would be viewed through the lens of your supervisor. Of course it is much more difficult with toxic PIs and sometimes outright impossible. However, it is important not to view oneself as entirely passive in that situation. Even unintentionally (and sometimes even by being passive) one can deteriorate a bad situation.

The interesting bit is that while the depot effect was assumed to be true, there actually has not been any systematic study on depot effects (and as Phil mentioned, adjuvants are now routinely used). Rather since the development was rather early on already focused on intramuscular (or subcutaneous) injections and shown to be effective at that, that intravenous injection never really became relevant. As such there were at best only case studies of accidental IV injection, but really no systematic investigation of possible issues.

So, some of the things you mentioned that folks describe you (e.g. unmentorable, or getting accused of falsifying data) are quite significant red flags. But without knowing context, it is difficult to tell whether you got in some unfavourable situations or whether you contributed to them. The other part are unrealistic goals. As a PhD candidate it is not typical in biomedical research to expect publications within the first year, unless you finish up someone else's work. And even then the PI typically rewrites the whole thing, anyway. From the perspective of a mentor your description of the events mostly show that things happened and that you are upset with it. But it does not tell anything about your suitability for a science career (or any other career for that matter). From my perspective you have to deal with two things. The first is the emotional aspects, which Koti covered. The second is figuring out the tangibles that got you these evaluations. How was your relationship to your supervisor, what were the issues, how did you address them, were you dismissive... and so on.

Looking at the harmful level based on the MSDS sheets should give you a good idea about dangerous level. While I loathe to give advice on that matter over the internet (and you really should take about that with your health and safety officer and/or your supervisor), a simple estimate should provide you some insight whether you have approached that levels. You should also inform yourself on the bioaccumulative features of the chemicals that you are handling (though suffice to say, it is eliminated very quickly). There are two general comments though. A) work with hazardous substances should always be under the hood, and disposal of volatile substance have to be sealed. I.e. even if not immediately harmful, you should always limit exposure and your safety officer should let you know how to dispose them in a safe way. The second which seems off is that you put them into a biohazard bag. That would imply that you are also working with biohazards that come into contact with your cotton?

OK, that seems to be arbitrarily stupid. Edit: talking about stupid: Considering all the effort surrounding elections and the fact that campaigning for the next election in the US pretty much seem start the day after any given election, this looks like a rather huge oversight...

I agree with that and am uncomfortable with some plans to have a school-to-company approach. Often also industry likes to outsource on-site training as much as possible to the unis and some jump on that,. While it provides short-term benefits for the graduates, I do not think that it is a good strategy overall for a variety of reasons, including what you mentioned.

That is wrong. Fibrous tissue is generally not consist of cells, but are part of the extracellular matrix (e.g. secreted collagen). For production of collagen and other fibers either stem cell type have to undergo a bit of differentiation. Fibrous tissues are also not scars per se they are structural and part of your normal body. Some types of scar tissue are characterized by a higher deposition than the normal amount of fibrous proteins and/ or different ordering so that the tissue looks different. That seems to be related to higher interference by the immune system to some degree. I.e. with a mature immune system, scarring increases.

Chemotaxis typically refers to a hard-wired response of sorts. It can be based on chemical pathways within a cell, but can also be a multicellular system that forms a circuit that reacts autonomously to chemical input. It is not usually used in contexts of e.g. learned avoidance. Thus it can be applied to multicellular organism but usually only in the context of hard-wired responses.

I can agree with that. In systems where students accrue debt to obtain their education, it is easy to see why it becomes a fiscal decision. It does colour interaction to a large bit. There are increasing student services on campus and many folks (myself included) try to highlight the transferable skills (i.e. not the ability to memorize the right answers, but rather the ability to organize info and oneself, for example). You are also correct that despite all that there are few areas where a career path is obvious. Many students start thinking about their career only when they get closer to their degree and just assume life will continue in predictable paths, which it does not. As a consequence this: Is something I have been struggling with. Universities are set up to kind of provide both, students are obviously more interested in one or the other and I do feel that our curricula does is a bit ambivalent to reach either goal (but then I often think strongly in terms of optimization and may therefore overthink this issue). On a broader issue, I do think that a better educated public is a common good and it should not be on the individual to get enough money to afford one. With fiscal issues at the forefront it also leads to the rise of exploitative educational outlets. But even among proper universities some leadership are thinking of the institution as a business rather than a public service, a stance which I do not quite agree with.

Also note that after Wakefield many folks looked into the link and if one does not cherry pick or falsify data, the association vanishes. Meanwhile, lack of immunization has brought back diseases that were on the brink of vanishing, and children pay the price for that.

I do not have a link, I just have the paper here. However, it should be easy enough to find. The full reference is Sejvar et al 2016 , Clin Inf Disease 15;63(6):737-745. If you are unfamiliar with literature references, you still should be able to get a hit by just googling it (or use scholar.google.com for example.

Your interjections are hard to read, but take a look at the citation I provided earlier. There the case for 2014 is listed by week.

Not really, if you consider the number of children sitting in back seats. In 2014 there were about 70 million children and the vast majority would have been in cars at some point. That would have an incidence rate of roughly 1.7 per million cases. Even if we think that only half of them were in cars and if we assume that indeed AC exposure would be a strong factor it would still be just 3.4 cases per million. In almost all situations these would be considered rare cases. The rate for non-symptomtic enterovirus infection is much, much higher for example. To provide other reference values, in the US in the same year about 16,000 children were shot and about 2,500 died from gunshot. Perhaps on would not consider it a rare event at this point, but consider this, in the last few years about 50 people (i.e. just a tad less than half of the AFM events in children) were shot annually accidentally by toddlers. I think at this point it is really fair to declare the range as "rare". But regardless of how to classify it, the issue is still that for a proper epidemiological analysis we simply do not have enough data points. even if it is 300 across the nation, creates too many variable to reasonably handle and too low statistical power to get some conclusion. If the use of ACs was a rare event, it would be easier to figure out. Even worse, even if we assume that AC use would be reasonable cause until end of September, we still have got almost half of the cases (~55) between October and December. So AC use could reasonably only be associated with half of the already low data set. My overall point is that you may look at things too specifically, the link is not terribly strong and, perhaps even more complicated, if it is really viral, then there is no good way an AC can contribute. Other factors including genetics, previous diseases or perhaps even diet could equally or more play a role and should not be overlooked in favour of one particular pet theory. While it may be worthwhile to look into other agents including mold, there is little to support that at this point. For example, in Florida the frequency is the same as in Arizona (one case each). Ultimately, one key aspect is identifying the actual causative agent, which will require more biomolecular work.

The issue is that e.g for a peak year in 2014 there were only 120 pediatric cases. I.e. we only have 120 data points to look into and as such statistical associations will be very, very weak. That being said, the latest lit implies that enterovirus D68 may yet be the strongest link and that AFM is a rare (as many more do get infected without these symptoms), but severe symptomatic manifestation in certain susceptible folks.

I stand corrected, then. Also got Georgia wrong in as IIRC it would trigger a run-off and not a recount if they are close enough. Two errors. Yay. Edit: there is an interesting article showing how voting blocks have changed over the years.

So eight days in it looks like Dems picked up 34 seats (with eight to go) , gained six governor seats (two being in recount) . In the senate race after picking up AZ, they will probably lose two seats. From there, I am curious, do you still consider it overall disappointing as whole? Or does it feel disappointing because high-profile races were lost?

Stochastically it may be actually be correct as generally there is only one (or a limited set of complementary) true findings and virtually unlimited false ones. I would maintain that the overall methodology works better in practice than alternative approaches, especially those that are resistant to weeding out incorrect assumptions.

If you have a glass and pour water into it until it runs over and see a puddle on the table, do you stop pouring or do you continue because "drinking glasses hold water, everyone knows this"? We see the equivalent of the puddle in atmospheric CO2 measurements.